IBM to offer on-demand HPC cloud

IBM looks to provide effective delivery of HPC resources in a hybrid cloud environment
Written by David Chernicoff, Contributor

There was a time when high-performance computing (HPC), then almost exclusively referred to as “supercomputing,” was strictly the domain of a well-defined set of vertical markets, each with a specific set of needs that demanded the best computing performance that money could buy. But things have changed, and HPC is now a potential game-changer in mainstream business, especially when dealing with the huge amounts of data that can be collected and analyzed for almost any business model.

Still at the high end of the cost model for computing resources, HOC got a significant boost when vendors began offering HPC cloud services, allowing business users to get their feet wet with the gigantic upfront costs usually associated with HPC. And in its latest announcement, IBM has combined the latest in cloud technology, computer hardware, and HPC software to offer customers custom HPC solutions as an on-demand service.

The IBM Platform Computing Cloud is a hybrid solution targeted at customers who will need to burst their computing performance by offloading part of it outside of their own datacenter to the IBM cloud service. The service combines IBM's powerful HPC computing hardware with Software-as-a-Service (SaaS) middleware tools, many of which were acquired in the acquisition of Platform Computing in 2012. Platform Symphony is a parallel computing and application grid management software solution, while Platform LSF is the management platform for distributed HPC computing.

The cloud-based solution offers a number of advantages to the customer over building the capabilities in house. The most obvious is the reduction in start-up costs; IBM will build the cluster-to-customer specifications and then provide it as a service on an hourly or monthly fee basis. Management and support for the cloud cluster is also included in the fees, reducing the customer’s reliance on developing or hiring the in-house skills necessary to successfully optimize and run an HPC cluster.

Support on their cloud platforms is available for IBM AIX, HP B11.31, and multiple versions of Linux and Microsoft Windows Server.


Editorial standards